What is Predictive Run to Failure?
John Todd, Sr. Business Consultant/Product Researcher, Total Resource Management (TRM)
Posted 10/31/2024 🎃 Happy Halloween!
“How dare you mix the modern word predictive with the ancient concept of running equipment to failure! Those are on opposite ends of the maintenance maturity spectrum! See the graphic below… what is wrong with you!”
No one has ever guaranteed that your tried-and-true paradigms will stay in place forever. Things change over time, becoming either more sophisticated, or simpler.
Sure, as the diagram suggests, one would not typically run the more critical equipment under your purview to failure. For those devices, a more formal and yes, predictive, approach would make good sense. But…
What does RTF really mean?
Our perception of RTF is a piece of equipment that is run flat-out to the edge of its operating envelop with little to no attention paid until it grinds itself to death. Then, we simply remove it, dispose of it (properly of course) and replace it with a new one. Restart the system and move to the next cruller.
The reality is more along the lines of the equipment running in the background with perhaps the occasional preventive maintenance such as lubrication. Yet, in the end the equipment is removed/replaced when it ceases performing to expectations. It simply is not worth paying more attention because of the low cost of the replacement.
Maybe the equipment is then sent to be rebuilt, returned in a refurbished condition, then placed on a shelf for the next replacement. This may make financial sense. Not everything needs to be bright and shiny out of the box to suit the needs of the production line.
No matter the degree of RTF you are practicing, it is just as valid as any other maintenance approach depending on the economics of the equipment in the context of its operation.
What does “predictive” really mean?
The first image that might come to mind is probably along the lines of hocus-pocus and a bunch of AI marketing hoo-haw. An excellent start… you are skeptical. Good!
Let’s go back to basic statistics class for a moment. You all remember statistical process control (SPC), right? Upper limits… lower limits… any data point outside of those bounds is suspected to be an anomaly and should be investigated, not ignored, right? Bringing back nightmares yet? (“Should I use Student’s-T or maybe Weibull for this data set? Oh no, the space bus to Toledo left without me and my pet cactus!”)
Detecting anomalies is a key part of starting to predict what the future of your equipment looks like. We expect the equipment to operate within a range or envelope of performance. If we begin to see excursions outside the bounds, then we know something is starting to brew.
Yes, these anomalies can be a result of startups/shutdowns and external factors such as weather patterns, but over time there is a picture of what is “normal” and what is not. Given the right tools, you can account for these times of expected anomalies, focusing on the “real,” ones.
Being able to use equipment telemetry to paint the picture of the now and the future is a powerful tool for sure. It comes down to available data and time proven statistical models and methods that can expose brewing troubles with your equipment. Maybe your early warning is a few days or perhaps even months before, indicating that a failure pattern is developing… perhaps well before any human could detect. That might be quite valuable.
Putting the two ends together – Predictive Run to Failure
Now that you have calmed down a little, we can reason together and see how this idea might be of use.
Let’s go back to that hapless piece of equipment that is grinding away out on the production line. You know that it will fail at some point in time and need to be replaced. You also might also know approximately when that might be. Historically you can state with confidence that about once a year it needs attention.
What if, however, you were able to get a clearer indication that the equipment was beginning to fail? Would a few days, weeks, months advance warning be of value? Would that help with planning for the replacement (aka downtime)? Would the total/actual cost to production be reduced if you had early(er) warning? Would early warning make scheduling contractors or procurement of replacement items easier? What if you are seeing earlier and earlier failures of the equipment… would that be important knowledge to investigate what might be causing that?
Yes, the RTF equipment would need to be worth the effort to establish a predictive system. If the whole annual replacement activity costs $100 then there is no sense in doing anything different. But go back to the early warning part… that brings efficiency to the process and that can have big value.
There are reasonable cost solutions such as Aspen Mtell that are designed to not only detect anomalies, but also catch developing failure signatures… hidden gems in the volume of equipment telemetry. Given a steady stream of equipment data and your historical knowledge, solutions such as this can quickly have value, no matter what your approach to maintenance.
Wrap up
That wasn’t so bad, was it? Mixing opposite ends of a spectrum together to build something better. The value is in the early warning. In the end your equipment is still RTF and replaced. That does not change. But you now know WHEN the failure is about to occur and can begin to plan for the replacement. This is not hocus-pocus but rather using the data you are collecting anyway and leveraging proven statistical methods to give you insight.
John Q. Todd
John Q. Todd has nearly 30 years of business and technical experience in the Project Management, Process development/improvement, Quality/ISO/CMMI Management, Technical Training, Reliability Engineering, Maintenance, Application development, Risk Management, & Enterprise Asset Management fields. His experience includes work as a Reliability Engineer & RCM implementer for NASA/JPL Deep Space Network, as well as numerous customer projects and consulting activities as a reliability and spares analysis expert. He is a Sr. Business Consultant and Product Researcher with Total Resource Management, an an IBM Gold Business Partner – focused on the market-leading EAM solution, Maximo, specializes in improving asset and operational performance by delivering strategic consulting services with world class functional and technical expertise.